Serpent Monte Carlo Neutron Transport Code

Transcription

1 Serpent Monte Carlo Neutron Transport Code NEA Expert Group on Advanced Monte Carlo Techniques, Meeting September Jaakko Leppänen / Tuomas Viitanen VTT Technical Research Centre of Finland

2 Outline Brief history Methods and capabilities: Physics in a nutshell Depletion routine Group constant generation HTGR capabilities Typical applications Project organization and funding User community Serpent 2 2

3 3 Brief history The development of Serpent started in September 2004, under the working title Probabilistic Scattering Game, or PSG Topic for a D.Sc. thesis (completed in 2007)1 Built-in depletion routines and public distribution in Serpent can be characterized as a three-dimensional continuousenergy Monte Carlo reactor physics code, that is specialized but not limited to lattice physics applications Re-writing of source code (Serpent 2) started in October ) J. Leppänen, Development of a New Monte Carlo Reactor Physics Code. D.Sc. Thesis, Helsinki University of Technology, 2007.

4 4 Overview Three-dimensional geometry model based on surfaces, cells and universes (similar to MCNP and KENO-VI) Additional geometry types for reactor physics calculations: Nests : cylindrical fuel pin and spherical particle 3 basic lattice types with variations Explicit model for stochastic HTGR geometries Criticality and external source modes for neutron transport simulation (Serpent 2 also tracks photons)

5 5 Physics in a nutshell (Serpent 1) The laws of physics are based on ACE format data libraries, with a few additional tricks: Cross sections reconstructed on a unionized energy grid speed-up in calculation but increase in memory demand2 DBRC for resonant scattering kernel Built-in Doppler broadening preprocessor routine3 Geometry routine based on combination of surface-tracking and the Woodcock delta-tracking method4 well-suited for reactor calculations but poor performance in detector modeling J. Leppänen. "Two Practical Methods for Unionized Energy Grid Construction in Continuous-Energy Monte Carlo Neutron Transport Calculation." Ann. Nucl. Energy, 36 (2009) ) T. Viitanen. "Implementing a Doppler-Preprocessor of Cross Section Libraries in Reactor Physics Code Serpent." M.Sc. Thesis, Helsinki University of Technology (2009). 4) J. Leppänen. "Performance of Woodcock Delta-Tracking in Lattice Physics Applications Using the Serpent Monte Carlo Reactor Physics Burnup Calculation Code." Ann. Nucl. Energy, 37 (2010) )

6 6 Depletion routine (Serpent 1) Built-in depletion routines for tracking nuclide concentrations as the fuel is burnt: 5) Radioactive decay and fission yield data read from ENDF format files Decay / transmutation chains generated automatically without user effort Bateman depletion equations solved using the CRAM matrix exponential method, developed at VTT specifically for Serpent5 Typical depletion calculation tracks the concentrations of ~1500 nuclides Optional equilibrium Xe-135 calculation M. Pusa and J. Leppänen,. "Computing the Matrix Exponential in Burnup Calculations." Nucl. Sci. Eng., 164 (2010)

7 7 Group constant generation (Serpent 1) Serpent has the capability to automatically produce all input parameters needed for few-group nodal diffusion calculations: Homogenized multi-group cross sections and diffusion coefficients Scattering matrices Point-kinetic parameters (forward flux-weighed) Assembly discontinuity factors Pin-wise power distributions Effective delayed neutron fractions ( Meulekamp's method ) A semi-deterministic method for homogenizing group constants in critical spectrum (B1 fundamental mode calculation) 6 E. Fridman and J. Leppänen. "On the Use of the Serpent Monte Carlo Code for Few-group Cross Section Generation." Ann. Nucl. Energy, 38 (2011) )

8 8 HTGR calculation (Serpent 1) The delta-tracking method works exceptionally well in HTGR particle fuels Explicit model for stochastic geometries: Coordinates of spherical objects are read from a separate input file Works at several levels (TRISO particles in graphite / fuel pebbles in pebble-bed core) Tested in full-core pebble-bed reactor calculations (ASTRA experiment), no major increase in calculation time compared to a regular-lattice calculation7 Pebble-wise power distribution printed in a separate output file H. Suikkanen et al., 2010 "An Approach for Detailed Reactor Physics Modelling of Randomly Packed Pebble Beds." In Proc. HTR Prague, Czech Republic, Oct )

9 9 Typical applications Serpent is specifically developed for: Generation of homogenized multi-group constants Fuel cycle studies and depletion calculations at the fuel assembly level But also: Research reactor applications Validation of deterministic transport codes Educational purposes Examples of user applications:

10 10 Project organization and funding The Serpent team at VTT: Jaakko Leppänen Maria Pusa Tuomas Viitanen Ville Valtavirta (starting 10/2012) Funding from three projects: SAFIR-2014 (National safety research programme): Four-year project (KÄÄRME) started in 2011, total budget of 955 k HPMC (EU): Three-year project started in 2011, total budget of 209 k NUMPS (Academy of Finland): Four-year project, started in 2012, total budget 1.3 M

11 11 User community Serpent is used in 76 organizations in 27 countries around the world, number of registered users is around 180 Serpent website and discussion forum: Distribution of Source code by OECD/NEA Data Bank and RSICC Updates are distributed to registered users by (current update , distributed August 16, 2012) Second International User Group Meeting in Madrid, Spain, this week (September 19-21, 2012).

12 The Serpent code user community 12

13 13 Serpent 2 The main focus at the moment is the development of Serpent 2, started in October 2012: Removing memory limitations in Serpent 1 and extending burnup capability to ~100,000 depletion zones (done) Parallelization using hybrid OpenMP / MPI techniques insted of just MPI (done) Photon and coupled photon / neutron transport routines mainly for the purpose of reactor physics and shielding applications (started) Development of variance reduction techniques (not yet started) Multi-physics applications (started, major focus from 2013 on) Serpent 2 beta testing started in January 2012 (~50 users) Public distribution is scheduled for